A minimal symbolic assembler for MS-DOS 2.0 compatible systems.

• Nested INCLUDEs—as in, an INCLUDEd file can INCLUDE other files which in turn can INCLUDE other files...etc. How deep it goes is determined only by the size of the buffer.¹
• Covers every possible 8086 and 8087² instruction encoding, including undocumented instructions.
• Doesn't give a hoot about code formatting/spacing.
• Supports undefined data and the basic directives EVEN, *EQU³, END, and ORG.
• Accepts binary, quaternary, octal, hexadecimal, and decimal numeric constants. Quaternary is in there⁴ because it was a debug thing that got left in. It takes up all of 2 lines (4 bytes) of extra code, oh no. not anymore it isn't
• Built-in VERY basic disassembler.
• Basic listing functionality.

• No macros.
• Segments schmegments!
• No OBJ files, relocations, segment trickery, external symbol definitions, basically anything that is remotely useful to the modern programmer who is into linking and modularity. And bloat.
• No algebra. It'll do the basic +/- arithmetic, but no fancy infix stuff or other operators.
• No MASM/Intel-style displacement voodoo with consigning displacement registers to separate sets of brackets (e.g. [BP][DI]) or appending brackets to variables BASIC-style to pretend play arrays (e.g. BYTE PTR BENZINO[BX+54]) or even weirder stuff like 40H[SI]. RA could potentially be spruced up to handle things like this but I deemed it hideous and bailed. Displacements entirely go inside brackets, like WORD PTR [NAPALONI+BP+SI-0BEEFH].⁵ I figured this is assembly, not bloody BASIC.
• Numbers may be 16 bits at most. DD/DQ/DT will reserve the proper amount of bytes, and any 16-bit value will be sign-extended to be stored, but there's no higher math beyond this. I chose to leave it at that because in practice, 16-bit assembly at its lowest level uses, well, only 16 bits. If anything, ADC DX,0 and SBB DX,0 are your friends.
• On a similar note, no floating point numeric constant handling/display. This is for the simple reason that I have absolutely 0 concept of how that would even work. I'm bad at math. I bring you instruction encodings, not Einsteinian theorems. There's probably someone out there more savvy than I who could implement it without breaking a sweat. I would not just be breaking a sweat; I'd be breaking every single braincell in the process.

• You use this program and its accompanying examples and utilities entirely at your own risk. I may not be held responsible for damage caused to your system through improper use of this program.
• This program is provided as is, with no warranties express or implied, and no development schedule. It is updated at my leisure and at my whim.
• I am not responsible for physical accidents, deaths, and other disasters caused by improper use of the program.